MOODZ BI-MODAL PI - VLF TRANSMIT CIRCUIT

[moodz]

So in other threads we have many types of detectors ... it would be nice if we had a general purpose uber transmitter circuit that would suit PI and VLF detector receivers.
I had in mind my own MAGPI detectors ( V4 and can be used on V3 with some modifications ). Also Carl's AMX CCPI project and there are probably many more out there that could use this cct also.
This cct allows the use of prexisting DD or OO coils designs and commercial and it has "proper" bipolar current AND voltage waveforms.
It can switch from PI mode to VLF very quickly without changing coils.
The original CCPI circuit had monopolar flyback voltages. This cct divdes the peak flyback voltage in half and also swings pos / neg instead of monopolar.

note it has been bench tested and it works ( before the sim jockeys reach for their keyboards )

This particular configuration can achieve 2 amps peak to peak in PI mode using SIC mosfets ( 2.5 microseconds transition ) and 4 amps peak to peak in VLF mode .

The circuit is below ...

moodz

[moodz]

RX waveform in PI mode. ( straight off the coil ... no processing / filtering etc )



RX waveform in VLF mode ( again straight off the coil no filtering etc )

[moodz]

...oh yeh the TX waveform ... showing settling time with a 2 millisecond changeover between VLF and PI ( of course you can also just leave it on PI or VLF )

[Altra]

Hi Moodz. This looks like it has a lot of potential. What are you proposing for a practical pulse transformer?

[KingJL]

Hi Moodz. This looks like it has a lot of potential. What are you proposing for a practical pulse transformer?

Not Moodz, but... you could make your own. I have made (wound my own) some pulse transformers in the past. A possibility is winding on a Fair-Rite Toroid 5975000211 (Type 75, 9.74mm OD, 4.56mm ID, 3.34mm H), 9.75 T/19.5T AWG 30 (could be 10T/20T). The resistance would be ~0.06/ 0.12 ohms. You could go to a thinner wire, but AWG 30 is the smallest I work with.

[Altra]

Hi JL,
l remember making the gate drive transformers. Looking at the schematic, it looks like a high current transformer. Much like the deemon CCPi xformer?

[KingJL]

Looking at the schematic, it looks like a high current transformer. Much like the deemon CCPi xformer?

1.6A primary and secondary(the primary circuit and secondary circuit resistance is important).

[Altra]

Ok, Thanks.

[moodz]

Hi Moodz. This looks like it has a lot of potential. What are you proposing for a practical pulse transformer?

This circuit is the result of building it first as a POC ( proof of concept ) then trying to model it in LTSPICE.

The model is deficient as the actual transformer series resistances at DC are lower ( approx 0.1 ohms ) so the higher resistances in the model are due to skin / core losses. ( so need proper models to sim )

The transformer is a 2:1 to match impedance however you could lift or lower the ratio to adjust for Tilt ( I let you guys try that experiment ... too long to explain here ).
Tilt can be adjusted by varying R2 in the diagram above ( the value of R2 will also be dependant on your transformer ratio ... its alot easier to vary R2 than vary the turns ratio in real time )

The transformer in POC is a simple toroid wound with 0.8 mm magnet wire. Note also the output FETS of the CCPI pulse driver are paralleled with schottky diodes to lower the forward conduction voltage when using SICS.

The toroid in this case has about 100 turns ( double wire ) on the primary and 50 turns on the secondary.

It works really well for what it is .... you can adjust the peak to peak values in the output TX by simply adjusting the supply voltage ( eg 2 - 5 volts ) ... at 5 volts we are maxing the SIC breakdown voltage.

I will be winding a more "efficient" toroid to see if it makes any difference.

In the pic below I just used a mechanical switch to switch from PI (open ) to VLF (closed) ... it works fine.

It's a very simple circuit. as you can see from the picture. The test coil is a commercial Minelab DD 43cm. (note I used a 1 uF cap in the pic and 2k for the tilt adjust and the config is a little different in layout -- but fairly much the same performance as the sim.) the tx pulse frequency will determine what size cap you use ( another experiment for the dear reader )

[Altra]

Thanks for the detailed reply!

[wtrk]

Hello brother it looks good. Will you share pcb files for check and what distance do you think distance with minelab coin ?

[eclipse]

This looks as versatile as you can get. Thanks for posting moodz, this goes in my bucket list to try.

[moodz]

Hello brother it looks good. Will you share pcb files for check and what distance do you think distance with minelab coin ?

Thanks .. I will post some board info when I get the final tweaks done. Cant really tell what distance on a coin ... as this is only the TX .... no RX circuit.
Will let you know the RX test results shortly.

moodz

[moodz]

This looks as versatile as you can get. Thanks for posting moodz, this goes in my bucket list to try.

Thanks Eclipse ...

I have rewound a more efficient transformer and now get 2 amps peak to peak in PI mode for 500ma @ 5 volt supply = 2.4 watts ( the original transformer was 1 amp at 5 volts = 5 watts for the same PI pulse )
Can still run the circuit at lower voltage (eg 1 amp peak to peak in PI ) for 1.5 watts consumption.

Also here is the FIRST evidence of discrim on the new circuit ... This is straight off the RX coil of the Minelab DD .. no amps / filters etc.

Blue = No target
White = rusty iron sledge hammer head
Green = 10 x 10 cm aluminium foil

Targets were 10cm from coil.
There is clear pivot of the ferrous vs non ferrous visible in the waveform. The pivot is at the graticle center line.

PS ... I dispensed with tilt control.. you dont need it if the losses in the transformer are minimised. The result below has no tilt control.

moodz

[green]

Hi moodz
Trying to understand your scope pictures. What I'm seeing, maybe all wrong. No target trace similar to target traces. Would be at least 7us before signal could be looked at if coil signal was amplified. Was thinking one of the advantages of CC Tx is early sampling.
Including some scope traces with normal Tx, can start looking at signal about 4us after coil off. No target not the same as bigger targets.